Recent Advances in Artificial neural networks for EM parameterized modeling and optimization

Abstract

This paper reviews the recent advances in artificial neural networks (ANN) for electromagnetic (EM) parameterized modeling and optimization. As an advanced ANN-based EM parameterized modeling and optimization technique, the neurotransfer function (neuro-TF) is discussed further in this paper. The trained neuro-TF parameterized models can be further used for EM design optimization with repetitive geometrical variations.